Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h Vo lume 3, Issue 3, Marc h -2012 1

ISSN 2229-5518

Dynamic Web Service Discovery Model Based on Artificial Neural Network with QoS Support

Shamim Ahmed, Momotaz Begum, Fazlul Hasan Siddiqui, and Mohammod Abul Kashem

Abs tractThe Universal Description, Discovery and Integration (UDDI) registries do not have the ability to publish the QoS inf ormation, and the authenticity of the advertised QoS inf ormation available elsew here may be questionable. We aim to ref ine the discovery process through designing a new f ramew ork that enhances retrieval algorithms by combining syntactic and semantic matching of services w ith QoS. We propose a model of Artif icial Neural Netw ork (ANN) w ith Quality of Services (QoS) based Web services discovery that combines an ANN based intelligent search and an augmented UDDI registry to publish the QoS inf ormation and a reputation manager to assign reputation scores to the services based on customer f eedback of their performance. We develop a service matching, ranking and selection algorithm that f inds a set of services that match the consumer’s requirements, ranks these services using their QoS inf ormation and reputation scores, and f inally returns the w eb service consumer based on the consumer’s pref erences in the service discovery request. Finally the w eb service discovery w ith QoS gives the most cost eff ective and suitable services as an output. The eff ectiveness of the system is improved by means of Artif icial Neural Netw ork w ith QoS.

Index TermsWeb Services Discovery, Quality of Services (QoS), Web Service Broker, Artif ic ial Neural Netw ork, UDDI.

—————————— ——————————

1 INTRODUCTION

HE proposed work is to provide an intelligent search for the relevant web service for the given set of requirements of the service consumer and based on the contextual in-
formation which is an input from the environment. The pro- posed intelligent search is planned at the consumer‘s end. This framework provides intelligent search to the consumer with the help of neural network. Neural network adjusts the weight of each node in the network by the trail and error method.
The consumer provides the set of inputs to system. In the pro- posed system, Input encoder will help to convert the user i n- put into input vector. Artificial Neural Network (ANN) a c- cepts only vector values. Inputs will be either real number or Boolean value. The resultant input vector is passed on to the

———— ——— ——— ——— ———

Shamim Ahmed has been serving as a lecturer, Department of Computer Science and Engineering (CSE), Dhaka International University (DIU), Dhaka, Bangladesh.He is also an M.Sc. in Engineering Student, Depart- ment of Computer Science and Engineering (CSE), Dhaka University of Engineering & Technology (DUET), Gazipur, Bangladesh. He got B.Sc. in engineering degree in CSE in the year of 2010 from DUET, Gazipur, Ba n- gladesh. Field of interest: Digital Image Processing, Artificial Neural Net- work, Artificial Intelligence & Visual Effects. E -mail: sha- mim.6feb@gmail.com

Momotaz Begum has been serving as a lecturer, Department of Computer Science and Engineering (CSE), Dhaka University of Engineering & Tech- nology (DUET), Gazipur, Bangladesh. Field of interest: Advanced Data- base System, Software Engineering . E-mail: momotaz03_duet@yahoo.com

Md. Fazlul Hasan Siddiqui has been serving as an Assistant Professor, Department of Computer Science and Engineering (CSE), Dhaka Universi- ty of Engineering & Technology (DUET), Gazipur, Bangladesh. Field of interest: Neural Network & Fuzzy System. E-mail: siddiqui.cse@gmail.com

Mohammod Abul Kashem has been serving as an Associate Professor and Head of the Department, Department of Computer Science and Eng i- neering (CSE), Dhaka University of Engineering & Technology (DUET),

Gazipur, Bangladesh. Field of interest: Speech Signal Processing. E -mail:

drkashem11@duet.ac.bd
ANN. Number of input and output nodes are defined by the designer of the network. But the number of hidden nodes and layers are not dependent to designer of the network. Normally one hidden layer is better to reduce the complexity of net- work. The weight is adjusted in network based on trail and error method. It will get experience and adjust the weight for each node. First ANN gives detailed possible information about various types of services. Then ANN produces the ou t- put according to that information as vector form. The vector contains the suggestion for desired services. This suggestion is passed to the Web Services Discovery with QoS for finding suitable service.
Service Oriented Architecture (SOA) is an approach to build distributed systems that deliver application functionality as services which are language and platform independent. A Web service is a technology that rea lizes the SOA. The current Web services architecture encompasses three roles: Web ser- vice provider, Web service consumer and Universal Descri p- tion Discovery and Integration (UDDI) [1]. Web service pro- viders use the Web Services Description Language (WSDL) [2] to describe the services they provide and how to invoke them. The service providers then register their services in a public service registry using UDDI. Application programs discover services in the registry and obtain a URL for the WSDL file that describes the service. Then, the applications can invoke the services using the XML-based Simple Object Access Proto- col (SOAP) [6] in either asynchronous messaging or Remote Procedure call (RPC) mode.
Finding the suitable service in the UDDI registry that satisfies the user needs or goals (Service Discovery) is the major prob- lem. In our proposed Web services discovery model, we ex- tend the traditional Web service model consisting of a service provider, a service consumer and a UDDI to include a Web service QoS certifier and a reputation manger, and use an

IJSER © 201 2

http :// www.ijser.org

Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h Vo lume 3, Issue 3 , Marc h -2012 2

ISSN 2229-5518

augmented UDDI that contains QoS information to allow QoS - based service discovery. The proposed new registry differs from the current UDDI model by having information about the functional description of the Web service as well as its as- sociated quality of service registered in the repository. The certifier verifies the claims of quality of service for a Web ser- vice before its registration. The Web Service Broker (WSB) consists of a service consumer, a UDDI registry and a reputa- tion manager and helps to discover Web services that satisfy the consumer‘s functional, QoS and reputation requirements. The reputation manager collects and processes service ratings from consumer, stores service reputation scores in a Web Ser- vice Storage (WSS) like a Rating Database, and provides the scores. We develop a service matching, ranking and selection algorithm based on a matching algorithm proposed by Max- imilien and Singh [3]. Our algorithm finds a set of services that match the consumer‘s requirements, ranks these services using their QoS information and reputation scores, and finally re- turns the web service consumer based on the consumer‘s pre- ferences in the service discovery request. Finally the web ser- vice discovery with QoS gives the most cost effective and sui t- able services as an output.
Neural Network is a machine learning technique. It is an in- terconnected group of artificial neurons uses a mathematical or computational model for information processing based on a connectionist approach to computation. Due to their parallel computing nature of neurons, it can perform computations at a higher rate compared with classical method. Neural network has adaptive nature. Due to this adaptive nature, system will adapt to the environment and input. It will produce the ou t- put depends upon its input. The output of one node will be the input of another node and the final result or output d e- pends on the complex interaction of all nodes.

2 RELATED WORKS

Several Web services may share similar functionalities, but possess different non-functional properties. When discovering Web services, it is essential to take into consideration fun c- tional and non-functional properties in order to render an ef- fective and reliable selection process. A number of research efforts have studied either QoS-based service discovery or reputation management systems. Invoking a low quality ser- vice in the system could affect the overall performance of the system, among the basic QoS factors are service performance (throughput, response time, latency, transaction time), viabil i- ty, accessibility, reliability, scalability, exception handling, execution cost, reputation, regulatory, accuracy, integrity, in- teroperability, security (authentication, authorization, confi- dentiality, traceability, data encryption, non -repudiation), pri- vacy, network-based factors (network delay, delay variation, packet loss), etc. [4]. Assuring the quality of the selected Web services was discussed in many proposals [4][7][8]. In [4] Ran extend the traditional service discovery model with a new role called a Certifier, in addition to the existing three roles of Ser- vice Provider, Service Consumer and UDDI[1] R egistry. The
Certifier verifies the advertised QoS of a Web service before its registration. The consumer can also verify the advertised QoS with the Certifier before binding to a Web service. This a p- proach prevents publishing invalid QoS claims during the registration phase, and help consumers to verify the QoS claims. Although this model incorporates QoS into the UDDI, it does not provide a matching and ranking algorithm, nor does it integrate consumer feedback into service discovery process. QoS can be used to select and rank the Web services by extending standard service oriented architecture (SOA) [10]. In this architecture, the Web service is selected by match- ing requested QoS property values against the potential Web service QoS property values [11].
Web service is the most interesting research area in Service Oriented Architecture (SOA). Finding a suitable web service at right time is a potential issue, needs to be addressed by the researchers. Available techniques for web service discovery are not proved to be efficient enough neither to discover the right and suitable services nor to find the services in time. The client will request the service according to their needs, situa- tion, and environment conditions. The services are discovered based upon requests of the clients and the contexts. Context aware service discovery aims to find the exact services based on the contexts at which the requests were given by the service consumers. Just-In-Time will provide the right service in right time to the right users. A framework for JIT-Oriented web ser- vice discovery using Neural Network. The effectiveness of the system is improved by means of neural network. The system will learn from its experience to predict user requirements and provide the services accordingly.

3 ARCHITECTURE OF HYBRID MODEL FOR

EFFECTIVE INTELLIG ENT SERVICE SEARC H

Let ‗W‘ be the proposed Intelligent Service Search system. It
can be defined as a set of elements {I, X, S, O}, where
I is the set of inputs (Ii=I1, I2, I3……….In), where
i=1,2,3……….n.
X is the input vector (Xi=X1, X2, X3……..Xn) for ANN, where
i=1,2,3……….n.
S is the set of suggestions (Si=S1, S2, S3,…….. Sk) ,where
i=1,2,3……….k. ‗S‘ value may be real number or Boolean.
O is the set of desires output services (Oi=O1, O2, O3….Om) from the proposed system to the consumer. It is the collection of desired services. where i=1,2,3……….m.
Let ‗IS‘ be the intelligent Search module. It can be represented
as-

IJSER © 201 2

http :// www.ijser.org

Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h Vo lume 3, Issue 3 , Marc h -2012 3

ISSN 2229-5518

Working principle of Intelligent Search System Effective Intelli- gent search system working can be divided into five stages. These stages are.

3.1 Req uest and Input St age

Consumer requests the services by providing input to intell i-
gent system to get desired services. The input for this stage can be defined as {I1,I2,I3…..In}. Overview of this stage can be defined as
Where,
‗X‘ is the input vector to the ANN. Where ‘n‘ is the number of
input elements.
‗H‘ is the Hidden nodes in artificial neural network. ‗i‘ is the number of nodes in 1st layer. ‗j‘ is the number of nodes in IIst layer. And so on.‘ k‘ is the number of layers in the network.
‗W‘ is the weight is passed to each node in network. Each weight has different value. Based upon input and output the weight can be modified.
Request and Input Stage => I= {I1, I2, I3…..In}.

3.2 Enc oding St age

Input encoder converts the user input into vector. Input of this
stage can be defined as {I1, I2, I3…..In} and produce the output as { X1, X2, X3……..Xn}.. The overall working of intelligent search is illustrated in figure 3. Overview of this stage can be defined as
Encoding Stage =>{ I1, I2, I3…..In } -> { X1, X2, X3……..Xn}.

3.3 Execution St age

The input vector is passed into ANN and it will adjust auto- matically based on the weight and produce the set of sugges-

Fig 1. Architecture of Hybrid Model f or Intelligent Service Search

‗S‘ is the output vector. It gives set of suggestions to the Ser- vice Manager. ‗m‘ is the number of elements in suggestion set. It will be either real number or Boolean values.
tions. The input for this stage can be defined as { X 1, X2, X3……..Xn } and construct output can be defin ed as { S1, S2, S3………Sk }. The overview of this stage can be defined as

IJSER © 201 2

http :// www.ijser.org

Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h Vo lume 3, Issue 3 , Marc h -2012 4

ISSN 2229-5518

Execution Stage =>{ X1, X2, X3…..Xn } -> { S1, S2, S3…Sk }.

3.4 Web Services D iscove ry w ith QoS Stage

This stage consists of UDDI [1] registry and Web Service Bro-
ker (WSB). The Web Service Broker (WSB) consists of Reputa- tion Manger, Web service QoS certifier, and Web Service Sto- rage (WSS). The web services discovery with QoS stage fetches the services with the help of suggestion vector based on ou t- put of ANN. The input for this stage be defined as { S1, S2, S3…Sk }. The certifier verifies the claims of quality of service for a Web service before its registration. The reputation man- ager collects and processes service ratings from consumer, stores service reputation scores in a Web Service Storage (WSS) like a Rating Database, and provides the scores when requested by web service consumer. The ANN acts as a broker between a service consumer, a UDDI registry and a reputation manager and helps to discover Web services that satisfy the consumer‘s functional, QoS and reputation requirements. The Output defined as { O1, O2, O3….Ok }. The overview of this stage can be defined as
Web Services Discovery with QoS Stage => { S1, S2, S3……Sk } -
> {O1, O2, O3…….O k }.

3.5 Out put and Terminat ion St ag e

Web Services Discovery with QoS Stage returns the more rel e- vant service { O1, O2, O3…….Ok } to the Service Consumer.
Output and Termination Stage =>list of services=> { O1, O2,
O3…..Ok }.

4 THE WEB SER VICES DISCO VERY WITH QOS

The Web Services Discovery with QoS consists of UDDI Regi- stry, Web Service Certifier with QoS, Reputation Manager and Web Service Storage (WSS).

4.1 UD DI Registr y and Web Service Cert ifier w ith QoS Web service provider needs to supply information about the company, the functional aspects of the provided service as requested be the current UDDI registry, as well as to supply quality of service information related to the proposed Web service. The claimed quality of service needs to be certified and registered in the repository [4]. Once the verification is passed successfully, the certification process is initiated. The certification process consists of issuing a certificate to the ser- vice provider. These certificate states that the offered QoS are conform to their descriptions. Th e Web service provider first needs to communicate its QoS claim to the Web service QoS certifier. The certifier checks the claims and either certifies or down grade the claim. The outcome is sent back to the provi d- er with certification identification information. A certificate is sent to the Web services provider and a copy is stored in the broker‘s database (WSS) identified by a certification Id for fu- ture use. A certificate includes information such as certificate number (certification Id), certificate issue date, and number of years in business, and services location. The certifier provides a set of Web services for any interested parties to access its repository about QoS claims for verification purposes. After the QoS certification been issued by the certifier, the supplier then registers with the UDDI registry with both functional description of the service and its associated certified quality of service information. The UDDI registry communicates with the certifier to check the existence of the certification. After successful checking, the registry then registers the service in its repository [4].

4.2 Re put ation Manager

The reputation manager collects feedback regarding the QoS of the Web services from the service consumers, calculates reputation scores, and updates these scores in the Rating DB. For this work, we assume that all ratings are available, objec- tive and valid. Service consumers provide a rating indicating the level of satisfaction with a service after each interaction with the service. A rating is simply an integer ranging from 1 to 10, where 10 means extreme satisfaction and 1 means ex- treme dissatisfaction [9]. Our service rating storage system is similar to the one proposed by Wishart et al. [5]. A local data- base contains the reputation information which consists of

IJSER © 201 2

Fig 2. Overall Working of Intelligent Search

Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h Vo lume 3, Issue 3 , Marc h -2012 5

ISSN 2229-5518


service ID, consumer ID, rating value and a timestamp. The } service key in the UDDI registry of the service is used as the } service ID, and the IP address of the consumer is used as the
consumer ID. Only the most recent rating by a customer for a service is stored in the table. New ratings from the same cus- tomers for the same service replace older rating [9].

4.3 Ser vice Matc hing, Ranking and Select ion Algor ithm When the discovery agent receives a discovery request, it ex- ecutes fMatch which returns a list of services LS1 that meet the functional requirements. If QoS requirements are specified, qosMatch is executed next on the set of services LS1 and it returns a subset of services LS2 that meet the QoS require- ments. selectServices always returns a list of M services to the customer where M denotes the maximum number of services to be returned as specified in the discovery request. If QoS requirements are not specified, selectServices returns M ran- domly selected services from LS1. If only one service satisfies the selection criteria, it returns this service to the customer [14].

Web services matching, ranking and selection algorithm:

1 findServices (functionRequirements, qosRequire- ments, repuRequirements,

maxNumServices)
{ // find services that meet the functional re- quirements

2 fMatches = fMatch (functionRequirements);

3 if QoS requirements specified

{ // match services with QoS information

4 qMatches = qosMatch (fMatches, qosR e-

quirements);
}

5 else

{ // select max number of services to be re- turned

6 return selectServices (fMatches, maxNum- Services, "random");

}

7 if reputation requirements specified

{ // matches with QoS and reputation in- formation

8 matches = reputationRank (qMatches, qo- sRequirements, repuRequirements);

// select max number of services to be re- turned

9 return selectServices (matches, maxNum- Services, "byQoS");

}

10 else

{ // matches with QoS information

11 matches = qosRank (qMatches, qosRe- quirements); // select max number of

services to be returned

12 return selectServices (matches, maxNum- Services, "byOverall");

Fig 3. Service matching, ranking and selection Algorithm

5 INTELLIGENT SEARCH IN TR AVEL

Let us consider the scenario; Mr. Sam, Manager for a small company residing in India has to attend conference in London. The mode of transport from India to London may be either through airlines or by ship. The proposed system, suggests the consumer about the convenient and comfortable service, using intelligent search. The system retrieves us er SSN, date, time and destination as input from user. The system learns with the help of neural network to suggest whether the customer can travel either by airlines or ship or any other mode of transport. According to the customer‘s scheduled conference date the system will check and infer that it is better to choose airlines instead of ship which may be a direct flight or connecting flight which ever is feasible to the customer. Based on his prior travel information the system will book his flights either in economy class or in business class. The proposed system re- trieves the services based on suggestion of ANN.
The output of ANN feed into the web service discovery with QoS. Finding the suitable service in the UDDI registry that satisfies the user needs or goals (Service Discovery) is the ma- jor problem. We extend the traditional Web service model consisting of a UDDI to include a Web service QoS certifier, Web Service Storage (WSS) and a reputation manger, and use an augmented UDDI that contains QoS information to allow QoS-based service discovery. The proposed new registry dif- fers from the current UDDI model by having information about the functional description of the Web service as well as its associated quality of service registered in the repository. The certifier verifies the claims of quality of service for a Web service before its registration. The Web Service Broker (WSB) consists of a service consumer, a UDDI registry and a reputa- tion manager and helps to discover Web services that satisfy the consumer‘s functional, QoS and reputation requirements. The reputation manager collects and processes service ratings from consumer, stores service reputation scores in a Web Ser- vice Storage (WSS) like a Rating Database, and provides the scores. Finally the web service discovery with QoS gives the most cost effective and suitable services.

Working Principle of Traveler Intelligent Search System

5.1.1 Req uesting St age (RS)

The Traveler (consumer) requests the service by providing set of inputs to the system. The input can be defined as {I1, I2, I3, I4…In}. That input set is {SSN, Distance, Date, Time}. SSN is the user Social Security Number. Distance is defined as dis- tance between source and destination of journey. Date and time attribute represent the Departure date and time. User input I1 is passed into Input Encoder. Overview of this stage is defined as

IJSER © 201 2

http :// www.ijser.org

Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h Vo lume 3, Issue 3 , Marc h -2012 6

ISSN 2229-5518

RS=>Input= I1= {SSN, Distance, Date, Time}

5.1.2 Enc oding St age (E S)

In encoding stage, user input is converted into vector. i.e. The user input set {SSN, Distance, Date, Time} is converted into input vector {X (SSN), X (Distance), X (Date), X (Time)} and produce as {13, 2000, 25, 9}. The resultant vectors either a real number or Boolean value. Because ANN accept either real numbers or Boolean values. The input vector I2 is passed into ANN for retrieving suggestion vector. I2 is defined as {13,
2000, 25, 9}. This stage can be defined as

ES =>I2= {X(SSN), X(Distance),X(Date),X(Time)} →{13, 2000,

25, 9}

5.1.3 Execution St age (E XS)

The input vector ‗X‘ be defined as {X1, X2, X3……..Xn} passed into ANN to get a suggestion set. First ANN gives detailed possible information about various types of services as preli m- inary output shown in table 1, table 2, and table 3.

Table 1. Illustrate the differences between various airlines according to available flight date, duration (including de- lay), and cost.

Airline

Available

Flight date

Duration (in-

clude delay)

Cost

Singapore

Airlines

27

9 hours

760$

Thai Air-

lines

26

10 hours

800$

American

Airlines

26

9 hours

710$

Table 2. Illustrate the differences among train, bus, and ship according to duration (including delay), and cost.

Transportation

Duration (in- clude delay)

Cost

Train

12 days

556$

Bus

23 days

400$

Ship

17 days

600$

Table 3. Illustrate the differences betwee n various hotels according to the cost.

Hotel

Cost

Skyline Hotel

60$ /night

Dream Night Hotel

71$/night

Sun and Moon Hotel

90$/night

The weight is adjusted automatically in ANN and retrieves the set of suggestion {S1, S2, S3………Sk} as output. i.e. input vec- tor {13, 2000, 25, 9} is passed into ANN and produce sugges- tion vector as output O1 is defined as {1, 0, 0, 0}. The values for suggestion set are {Flight, Ship, Train, Bus}. The value for each element will be either 1 or 0. The value ―1‖ suggests to use a particular travel commodity else 0. For this exa mple, the ANN
suggests to take fight travel. The output can be defined as {1,
0, 0, 0}. The output of ANN also is vector. It may be a single output or set. This suggestion set is passed to Web Service Discovery with Qos model. This interaction can be defined as:

Possible suggestion sets:

Transport => {Airline, Train, Bus, Ship} → {1, 0, 0, 0}

Airline => {Singapore Airlines, Thai Airlines, American Air-

lines} → {0, 0, 1}

Hotel => {Skyline Hotel, Dream Night Hotel, Sun and Moon

Hotel} → {1, 0, 0}

Train, Bus, Ship}→{1, 0, 0, 0}

5.1.4 Web Ser vice Discovery wit h Q oS St age and O utput

Stage:

The possible suggestion set is passed to Web Service Discovery with Qos model. The Certifier verifies the QoS claims from the Web service provider. An ANN helps finding services that meet the functional and QoS requirements specified by the consumers. With the assumption that the consumers provide non-malicious and mostly accurate QoS ratings to the reputation manager, these matched services are then ranked based on both their reputation scores generated by the reputation manager and their non - functional QoS attributes values. The top ranked services are r e- turned to the service consumers. Finally the Web service disco v- ery with QoS gives the most cost effective and suitable services as an output.

6 CONCLUSION

In this paper we have presented a new approach for Web Ser- vice discovery process. Due to the increasing populari ty of Web services technology and the potential of dynamic service discovery and integration, multiple service providers are now providing similar services. QoS is a decisive factor to distin- guish functionally similar Web services. We proposed a si m- ple yet novel approach to provide most cost effective and sui t- able Web service discovery is achieved using Artificial Neural Network with QoS. Context-awareness is utilized to find the perspective of users query. The system will discover the ser- vice based upon their consumer input. It helps to provide matching services to consumer by eliminating irrelevant ser- vices. Intelligent search is performed using Artificial Neural Network. Experience is fed as input to the intelligent search system. In scenario, Travel intelligent system, fetch desired services to traveler based upon their request. The paper presents an algorithm for effective service matching, ran king and selection. A mass of services is needed where we can test the performance of our system. For future work, we plan to extend QoS parameters to include information such as reliabi l- ity, fault rates and Security.

IJSER © 201 2

http :// www.ijser.org

Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h Vo lume 3, Issue 3 , Marc h -2012 7

ISSN 2229-5518

REFERENCES

[1] UDDI. The UDDI Technical White Paper. 2000. http://www.uddi.org/.

[2] R. Chinnici et al., ―Web Services D escription Language (WSDL) Ver-

sion 1.2‖, World Wide Web Consortium, www.w3.org/TR/wsdl12/,

2002.

[3] E.Maximilien and M.Singh, ―Toward Autonomic Web Services Trust and Selection‖. In Proc. of the 2nd Intl. conf. on Service Oriented Computing, New York City, USA, 2004, pp.212-221.

[4] S.Ran, ―A Model for Web Services Discovery with QoS‖. SIGEcom

Exchanges, Vol. 4(1), 2004. pp.1–10.

[5] R.Wishart, R.Robinson, J.Indulska, and A.Josang, ―Superstring Rep:

Reputation-enhanced Service Discovery‖. In Proc. of the 28th Austra-

lasian conf. on Computer Science, Vol. 38, 2005, pp.49-57.

[6] D.Box et al., ―Simple Object Access Prot ocol(SOAP) 1.1‖, W3C Note

08, World Wide Web Consortium, www.w3.org/TR/SOAP/, May

2000,

[7] A. Mani and A. Nagarajan, ―Understanding quality of service for

Web services‖, IBM paper, January 2002.

[8] Ying Li, Xiaochen DING, Ying CHEN, Dong LIU, Thomas LI, ―The

Framework Supporting QoS-enabled Web Services‖, IEEE Interna-

tional Conference on Web Services, Las Vegas, Nevada, USA, June

2003.

[9] Ziqiang Xu, Patrick Martin, W. Powely and F. Zulkernine, ―Reputa-

tion-Enhanced QoS-based Web service discovery‖, IEEE Internation-

al Conference on Web Services, 2007.

[10] M.Adel Serhani, Rachida Dssouli, Abdelhakim Haf idand Houari

Sahraoui, ―A QoS broker based architecture for efficient Web services

selection‖, Proceedings of the IEEE International Conference on Web

Services (ICWS'05), IEEE 2005.

[11] Demian Antony D‘Mello, V.S. Ananthanarayana and T.Shanthi, ―A

QoS broker based architecture for Dynamic Web services selection‖,

Proceedings of the Second Asia International Conference on Model-

ing & Simulation, 2008.

IJSER © 201 2

http :// www.ijser.org